Journal of Geodynamics 36 (2003) 423443

www.elsevier.com/locate/jog

Solar activity and earth rotation variability

R. Abarca del Rioa,*, D. Gambisb, D. Salsteinc,
P. Nelsonc, A. Daid
a
Departamento de Fisica de la Atmosfera y de los Oceanos (DEFAO), Facultad de Ciencias Fisicas y Matematicas,
Universidad de Concepcion, Casilla 160-C, Concepcion, Chile
b
International Earth Rotation Service (IERS), UMR8630, Observatoire de Paris, 61 Av de lObservatoire, Paris, France
c
Atmospheric and Environmental Research Inc, 131 Hartwell Avenue, Lexington, MA 02421, USA
d
National Center for Atmospheric Research, Boulder, CO 80303, USA

Abstract
The analysis of variability in Atmospheric Angular Momentum (AAM) and Length of day (LOD) of
Abarca del Rio et al. [Ann. Geophys. 18 (2000) 347] is extended to investigate a possible connection with
solar activity uctuations from interannual to secular time scales. The southern oscillation index and
records of sea surface temperature are used as proxy series in this analysis during the era prior to the
availability of AAM analyses. At interannual times scales, the variability in AAM and LOD agrees with
that in solar activity with regard to the decadal cycle in the stratospheric quasi biennial oscillation and
solar activity but whose phases are slowly shifting from one another with time, while the stratospheric
quasi biennial cycle agrees with the solar quasi biennial cycle, though led by 6 years. At decadal times scales,
AAM varies statistically with the solar decadal cycle over much of the last century since 19301940. The dec-
adal mode in AAM is suggested here to be generated by upward propagation of surface atmospheric modes,
from the surface throughout the troposphere through the stratosphere. Equatorial Sea Surface Temperature
(SST) variability may be considered a proxy index for AAM variability because of the relationship to the El
Nino/Southern Oscillation; its analysis over the last three centuries (17302000) and that of LOD since 1830
conrm the agreement found over the last part of the 20th century, as well as the general disagreement before.
# 2003 Elsevier Ltd. All rights reserved.

1. Introduction

The Sun drives the thermal balance of our planet and when modulated by orbit-based para-
meters, determines the seasonal weather cycles of our planet. Until recently Solar irradiance was

* Corresponding author. Fax: +56-41-204002.

E-mail address: rodrigo.abarcadelrio@udec.cl (R. Abarca del Rio).

0264-3707/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0264-3707(03)00060-7
424 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443

assumed to be invariant and thus the power intercepting a unit area at the Earths mean distance
from the sun was called the solar constant. Current satellite- and rocket-based instruments
demonstrate a variation in solar irradiance (Hoyt and Schatten, 1997) which is tied to the
approximate decadal cycle seen in sunspot numbers (Parker, 2000). This variability is relatively
small, 13 Wm2, corresponding to between 0.1 and 0.3% of the total solar irradiance reaching the
earth (Pap and Frohlich, 1999).
Recently, a number of studies have shown that from interannual to secular times scales,
meteorological and climatic data are correlated with solar variability (see reviews by Sadourny,
1994; Reid, 1995; Hoyt and Schatten, 1997; Lean and Rind, 1999; and Nesme-Ribes and Thuil-
lier, 2000). The earlier solar-climate link debate (Pittock, 1978, 1983) was revitalized with new
studies (Arnold and Robinson, 1998; Shindell et al., 1999) searching for a physical mechanism by
which small changes in solar output could be amplied in the Earths atmosphere. In fact, at
wavelengths not visible to humans, such as the Ultraviolet and Extreme Ultraviolet, the percen-
tage variability is much larger, changing by factors of 100 or more over time scales of minutes to
hours, and the Earth receives other types of energy such as outows of charged particles.
According to the recent Sun-Climate (2000) connection workshop held by NASA the research on
such links has been organized into general areas that we summarize here. (1) Direct forcing of
tropospheric climate by changes in the near UV, visible, and IR radiation. (2) Indirect forcing of
climate by solar induced changes in the stratosphere, such as through ozone interactions resulting
in temperature and other changes around the 30 hPa level, possibly modulating the Quasi-Bien-
nial oscillation (QBO). (3) Inuence of energetic particles, such as the galactic cosmic rays and
energetic electron precipitation modulated by the solar wind that can alter cloud cover as well as
induce upper and lower atmosphere couplings.
Recent results by Friis-Christensen and Lassen (1991), Crowley and Kim (1996), Diamantides
(1998), Kishcha et al. (1999), Mann et al. (2000) and Crowley (2001) as well as numerous refer-
ences in Lean and Rind (1999) have intensied the debate. Many of these investigations correlate
dierent indices of solar activity with hemispheric time series of surface temperature, suggesting
that up to 3050% of the interdecadal - century time scale variance in surface temperature could
be attributed to solar forcing. However, numerical experiments (Cubasch et al., 1997; Tett et al.,
1999; Bertrand et al., 1999; Meehl et al., 2003) show that although a combination of solar varia-
bility (interpreted nevertheless in view of reconstructed past solar activity indices), and volcanic
activities (e.g. Mann et al., 2000; Crowley, 2001) could have forced climate at secular time scales
up to roughly 1930, these can not be responsible for global warming, currently being observed
and reported about by IPCC (2001).
Because of the absence of consensus regarding physical processes relating solar activity to cli-
mate variations, this topic has caused much speculation (Pittock, 1978, 1983) and controversy.
Given results at secular and interdecadal time scales, one expects a similar relationship between
the powerful 11-year Schwabe cycle in solar activity and decadal times scales in climate indices,
though such results may be inconsistent; Nitta and Yoshimura (1993) as Parker et al. (1994),
analyzing global air (land + marine) temperatures, suggest that decadal variability in global
averaged air surface temperature anomalies and the 11-year solar cycle, though related, are
unrelated linearly before 1940. On the other hand, Lau and Weng (1995) and White et al. (1997),
found signicant solar-related signals in the equatorial sea surface temperature over the entire
20th century. Recently, though contested (Gierens and Ponater, 1999) it has been shown that
 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443 425

global cloud coverage (Svensmark and Friis-Christensen, 1997) contains a decadal cycle. Addi-
tionally, suggestions of the 11-year cycle appear in stratospheric parameters like geopotential
height, zonal winds and temperatures (Labitzke and van Loon, 1997) and noted in general for
some characteristics of the quasibiennial oscillation (Baldwin et al., 2001) over the last 40 years.
On this topic, much of the analyses has been performed with climate indexes at the surface in
the troposphere or in the stratosphere, or at individual geographically located parameters. In the
present study however, we will analyze an index representing and measuring the dynamic state of
the whole atmosphere, the atmospheric angular momentum (AAM). In fact, much of the historic
development of modern meteorology is connected with the study of how atmospheric momentum
is maintained locally or transported from one region to another and how it is exchanging
momentum with the oceans and solid Earth (Peixoto and Oort, 1992). Its conservation properties
are closely linked with those of energy conservation (White, 1989). It is therefore a fundamental
circulation index used to characterize the dynamic state of the general circulation of the atmo-
sphere, and climate (Peixoto and Oort, 1992). AAM is known to exhibit pronounced high fre-
quency (Schuh and Schmitz-Hubsch, 2000) and seasonal uctuations (Marcus et al., 1998;
Hopfner, 1998, 2001), interannual times scales associated with large scale ow anomalies (Rosen
et al., 1991; Dickey et al., 1992; Black et al., 1996; Chen et al., 1996; Dickey et al., 1999; Abarca
del Rio, 2000) and decadal oscillations (Abarca del Rio, 1997; Rosen and Salstein, 2000). In
addition, its knowledge is also important for space geodesy and interplanetary navigation, since
on all the above time scales, AAM is highly correlated with length of day (LOD) variations, a
measure of the Earths rotation rate (see review by Eubanks, 1993).
The present paper explores relations between solar activity (SA) described in indices like the
Wolf sunspot number (WSN), and atmospheric angular momentum (AAM), from interannual to
decadal times scales. This study covers two time scales; interannual and decadal. In the rst part,
we emphasize stratospheric quasi biennial (QB) variability as evidenced in AAM. We also inves-
tigate the triennial-quadrennial and six year oscillation (TQO and SYO respectively). In the sec-
ond section, we will concentrate our analysis in the decadal oscillation in AAM and LOD.

2. Data

2.1. Atmospheric angular momentum (AAM)

The axial AAM about the polar axis of a layer of the atmosphere, may be calculated following
Barnes et al. (1983):

2
R 3 Pt 
=2 2

AAM u cos2 @l @@p 1
g Ps
=2 0

where R is the radius of the Earth, g acceleration due to gravity and u zonal wind speed. The term
is integrated over all latitudes, , longitudes, l, and pressures, p (from 1000 hPa, near the surface,
to 10 hPa in the stratosphere. A second AAM term, related to changes in the mass distribution of
the atmosphere (surface pressure variations), plays a much smaller role in global AAM budget
(less than 5%) on most time scales (Eubanks, 1993), and is not accounted for in the following.
426 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443

Using the zonal wind elds from the recent NCEP reanalysis (Kalnay et al., 1996), we constructed
a 52-year long set of monthly global AAM from January 1949 to December 2000. The reanalysis
elds are available on 2.5 2.5 latitude-longitude grids over 17 layers from 1000 hPa to 10 hPa,
with ve levels over the stratosphere (70, 50, 30, 20 and 10 hPa). These elds allow computation
of AAM up to 10 hPa using Eq. 1. Recently, Rosen and Salstein (2000) showed that AAM data
issued from an atmospheric model (UKMO Hadley Center Model) forced by SST elds (GISST)
over 1870 to 1998 showed comparable variability to that of the reanalysis (over 19501998) in
seasonal to interannual times scales. With this in mind, we construct another AAM set with
data issued from similar model experiments (NCAR CCM3, Kiehl et al., 1998), forced by the
same sea surface temperature elds (GISST) over 18701997. From these models runs we con-
structed two ensemble averages of simulations of AAM, respectively for the UKMO and
NCAR models.
The AAM data can be expressed in millisecond (ms) of length of day (LOD) variability,
assuming that changes in AAM for the entire atmosphere are accompanied by equivalent changes
in the angular momentum of the Earth, through the following relation (Rosen and Salstein,
1983):
 
DLODms 1:68 1029 DAAM kg m2 =s 2

2.2. Length of day (LOD)

The LOD data are taken from the compilation of the International Earth Rotation Service
(IERS annual report, 1999) based on a combination of astrometric and space-geodetic methods,
from 1700 to 1830 at an annual resolution, from 1830 to 1949 seasonally, and from 1949 to
2000 monthly. From this data set we constructed a continuous series extending from 1700 to
2000. This data set matches well with the recent combined annual LOD data series from Gross
(2001).

2.3. Sea surface temperature (SST)

We will also analyze the equatorial SST variability from year 1730 up to 1980, taking advantage
of the recent annual reconstructed (marine + land) elds by Mann et al. (2000). For comparative
purposes we will also investigate the Kaplan et al. (1998) monthly elds of sea surface tempera-
ture from to 1856 to 2000. From these elds we constructed two main time series: Kaplan20 and
Mann5. The Kaplan20 time series is constructed by averaging globally (all oceans) from 20 S to
20 N the Kaplan et al. (1998) SST anomaly elds which are available monthly over 18552000 on
equal 5 latitudelongitude grids. The Mann5 time series is constructed by averaging globally (all
oceans) from 5 S to 5 N the anomaly elds of SST of the Mann et al. (2000) reconstruction,
which are available annually from 1730 to 1980 on 5 5 latitudelongitude grids. For both elds
we choose only grids which were complete over the entire time span and we constructed Pacic,
Atlantic, and Indian Ocean equatorial SST averages. In the case of the Kaplan elds, for com-
parative purposes, we also constructed 5 and 10 averages (Kaplan5 and Kaplan10 respectively).
Finally we will also use the global temperature (marine + land) time series by Mann et al. (2000),
extending from 1700 to 2000 hereinafter named GT.
 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443 427

2.4. Solar variability

For analytical purposes the intensity of the solar radio ux values at 2800 MHz (the 10.7 cm
radiation or F10.7 index), which measures the variable solar photon inputs into the atmosphere, is
likely the best suited for comparison with climate indices, though it starts in the 1950s (Gorney,
1990). For studies involving longer data series, the varying number of sunspots has been subject
of observation through several hundred years and may be regarded as reliable since 1750 (Eddy,
1976). The Wolf sunspot number (WSN) is highly correlated with Suns output of the radiation in
the extreme ultraviolet (EUV) wavelength band, and it varies identically with the 10.7 cm radia-
tion, with the eleven year period being its dominant oscillation (Gorney, 1990). In preliminary
analyses, we performed comparative analyses between the WSN and the F10.7 index over their
overlapping period (19502000) in all the frequency bands studied. We conrm that both series
are identical. Therefore we will use the Wolf sunspot numbers (the WSN) as a solar activity index
(herein after SA). Lets note here that the decadal oscillation in solar activity is also called the
Schwabe cycle following its discovery by Schwabe (1843).

3. Analyses

3.1. Interannual times scales

Interannual signals in AAM and LOD are dominated by a number of distinct bands (Abarca
del Rio et al., 2000). A 6-year oscillation can be noted (Abarca del Rio et al., 2000) in LOD and
up to 5 years, a pair of scales, related to ENSO emerges (Dickey et al., 1992; Black et al., 1996) in
both AAM and LOD, those at 35 years, and 23 years, also related to the quasi biennial oscil-
lation. We separate below the interannual uctuations into low and higher frequency (LF and HF
respectively) variability.

3.1.1. LF interannual variability

Association between ENSO and SA at interannual time scales have not been documented to
date. Results connecting the two processes appear to be conned to much longer times scales, the
8090 year Gleissberg cycle (Michaelsen, 1986), but such scales are far beyond those under con-
sideration here.
Given the close relationship documented between the southern oscillation index SOI and AAM
(see Abarca del Rio et al., 2000), and the availability of the SOI for over a century, since 1866, we
compared the interannual variability in this monthly SOI with the WSN series We computed the
spectra of each series, as well as co-spectra and squared coherence spectra (Fig. 1). We separated
the analyses into three spectral regions addressed in AGS, the six-year oscillation (SYO), the tri-
ennial-quadrennial oscillation (TQO) and the quasibiennial band (which is presented in the high
frequency section).

3.1.1.1. The six-year oscillation (SYO). Though neither the SOI nor the WSN peak near 6 years,
they do cohere strongly with each other at this frequency. Wavelet analysis of both series (Tor-
rence and Compo, 1998), are in phase in the 5.57 year band over 18662000 (not shown) though
428 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443

Fig. 1. Spectra of the solar activity (bold line), SOI (dotted line), co-spectra (grey dashed line) and squared coherency
(grey line) for the monthly time series for 18662000.

34 oscillations is often the limit of the phase agreement. Interestingly, the 6-year signal in LOD
noted by Vondrak (1977), and reconrmed with new space-geodetic data sets by Liao and Grei-
ner-Mai (1999) and Abarca del Rio et al. (2000), was shown by Djurovic and Paquet (1996) to
possess a signicant relationship with SA. Our analysis of LOD and SA over 1830 to 2000, by
both singular spectrum analysis (Vautard et al., 1992) with dierent embedding dimensions
varying from 8 to 20 years, and wavelet analysis conrms a present but transitory relationship.
The SA and LOD series peak in the 56 year and 67 year bands, respectively.

3.1.1.2. The triennial quadrennial oscillation (TQO). Although the SOI and WSN spectra in Fig. 1
are not signicantly related at the TQ time scale, we investigate the variability in AAM and LOD
at this scale, including a direct or lagged association with the decadal cycle in the solar activity.
Conrming spectra results in Fig. 1, in the 3.55 year band, SOI and SA do not present com-
parable variability (not shown). We used the amplitude of the TQ oscillation in SOI as a proxy
index of the possible modulation of the TQ band in AAM by the decadal-scale cycle in solar
activity (Fig. 2a). Such results also agree with earlier studies at interannual times scales in AAM
and LOD (Abarca del Rio et al., 2000), in which the TQ signal had a modulation of about 1315
years, associated with ENSO (Wang and Wang, 1996), and therefore dierent from the 11-year
period of the SA. The interannual and decadal variability in SA appears not to be related with the
TQO in neither SOI, AAM, nor LOD.

3.1.2. HF interannual variability: The Quasi Biennial Oscillation (QBO)

In contrast to lower frequency interannual signals, the quasi-biennial oscillation, particularly in
the stratosphere, appears to be related to both the phase of the decadal cycle in solar activity and
to the amplitude of a QBO signal itself. Ramanathan (1964, cited in Berson and Kulkarni, 1968),
found a break and phase shift in the QBO midway between times of sunspot maxims, based on
analysis of a series of total ozone data taken at Arosa (47 N) from 1939 to 1963. Later Berson
and Kulkarni (1968) analyzed a number of parameters including temperatures, total ozone and
wind, over a longer period, 1908 to 1965 and further conrming the existence of the QBO,
especially around the time of a solar minimum.
 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443 429

The period and amplitude of the solar cycle were noted to be related to rawinsonde-based zonal
wind observations during 19511979 by Quiroz (1981), and since other authors (Labitzke, 1987;
Labitzke and van Loon, 1990; Kodera, 1991; Naito and Hirota, 1997), have observed an even
stronger relationship when the data are organized according to the phase of the QBO. However,
such a stratifying approach has been criticized by Salby and Shea (1991) as introducing spurious
covariance through aliasing, which is not present in the unstratied data. Other critics cite a

Fig. 2. (a) Comparison of the Schwabe cycle (bold solid line) and the amplitude variability of the QB (dashed grey)
and TQ (solid line) in SOI. SA is dimensionless. (b) Comparison of the characteristics (period (solid line), amplitude(grey
dashed line) of the s-QBO in AAM and the inverted Schwabe cycle in SA (bold solid line). SA is dimensionless.
430 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443

relative short data length and the quality of data in the early part of the last century (Hamilton,
1990). Pessimistically, Hamilton (1998) provides a negative perspective on reconstructing a con-
tinuous and viable index of past stratospheric variability 1950.
Here we compare the quasi biennial variability in AAM (tropospheric and stratospheric) rst
with the decadal cycle in solar activity (subsection 3.1.2.1) and then with a comparable oscillation
in solar activity (subsection 3.1.2.2).

3.1.2.1. Association of the quasi biennial variability in AAM with the decadal cycle in solar activity.
The quasi biennial oscillation in AAM appears to originate from two separate processes, one in
the troposphere and linked with ENSO (the t-QBO), and the stratospheric (the s-QBO). Whereas
we were unable to note a link between the amplitude and phase of the t-QBO in AAM and the
decadal cycle in solar activity (Fig. 2a), we did determine, in contrast, a relationship, though
weak, with the s-QBO.
The evolution of the period and amplitude of the s-QBO are presented along with the inverted
decadal solar signal in Fig. 2b. Over much of the 50 years analyzed, a relationship ship exists
whose phase lag with the inverted decadal solar cycle appears to average 4 years, though it
vacillates from about 3 years in 1965 to 5 years in 1995. The variability in this lag might indicate
an origin for the decadal modulation other than solar activity, and furthermore the annual reso-
lution of earlier studies can partially mask the QBO period, particularly when segregated into the
east and west phases. To mitigate these issues in part, we were able to conrm the above rela-
tionship, with the vacillating time lags, using the expanded QBO index of Naujokat (1986) from
the mid-1950s between the 100 and 10 hPa levels. Following earlier studies separating the east
and west-directed phase, we found that part of the association between the decadal cycle in SA
and the period of s-QBO in AAM was due to a contraction of the east-to-west (through to peak)
changeover. When the contraction and expansion of the east to west change over is taken alone the
correlation appears to be amplied, but the association is also present in the west to east change over.

3.1.2.2. Association with the quasi biennial cycle in SA. Djurovic and Paquet (1993) noted the
relationship between a QB oscillation in green corona activity and one in Earth rotation. AAM
being the main contributor to LOD at these scales, it is therefore important to understand the
solar imprint, if there is one, within this band over the atmosphere.
Solar activity presents a quasi-biennial cycle, observed especially in sunspot numbers (Shapiro
and Ward, 1962; Akioka et al., 1987) as well as other solar activity indexes: solar variability (Bao
and Zhang, 1998; Benevolenskaya, 1998), neutrino ux (Sakurai, 1979, 1981), magnetic eld
(Riven and Obidko, 1992), radio ux at 10.7 cm, solar are rate, X-ray burst, sunspot area, solar
diameter variations in addition to the green corona activity noted above. It is a well-established
peak in solar activity although its amplitude is rather weak when compared to the 11-year cycle.
We will study its relationship respectively with the t-QBO and the s-QBO of AAM.

The t-QBO. The correlation of the quasi biennial oscillation in SA (QBO-SA) with the QBO of
AAM in the troposphere, is weak (correlation found is 0.41 SA at a time lead of 73 months) while
the correlation with the global AAM (troposphere and stratosphere) is 0.6, with the SA
leading (also by 73 months). An extended analysis of SA with SOI over 1866 to 2000 reveals
correlations of about 0.3, and moreover with the SOI leading SA by 50 months. However
 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443 431

when considering only the recent period since 1950, a similar level of correlation exists, though
with a reversal in the sense of the lag, with SA leading SOI.

The s-QBO. The signicant correlation found by Djurovic and Paquet (1993) between LOD
and the QBO-SA is explained by the relationship with the stratospheric contribution to AAM at
this band (the s-QBO). The s-QBO and the QBO-SA are correlated (0.54), with the solar cycle
leading the stratospheric angular momentum by 73 month (roughly 6 years), although the asso-
ciation increases over the last 30 years (0.6 with the sun leading by 73 months). If we extract the
quasi biennial variability through SSA (Vautard et al., 1992), a method to focus the data, over the
last 30 years the correlation is as high as 0.66 for LOD and 0.7 for s-QBO with the SA leading by
73 months. Using the 10.7 cm radio as alternative time series for SA, we obtained similar corre-
lation values, therefore conrming previous results with sunspots data. The comparison of QBO
in SA, though lagged by 73 months, with that in LOD, stratospheric AAM and tropospheric
AAM are shown in Fig. 3 (a, b and c respectively). It may be noted that when the QBO-SA signal
is at higher amplitude (19551960, 19751980, 19851990, 19952000), its phase is closer to that
in the stratospheric AAM lagged by 6 years (Fig. 3b). Elsewhere the phase is lost. Because the

Fig. 3. Comparison of the evolution of the quasi biennial (QB) cycle in SA (bold) with the quasi biennial (QB) cycle in
LOD (Fig. 3a), stratospheric (Fig. 3b) and tropospheric (3c) contributions (dashed line). All AAM series has been
lagged by minus 6 years towards the SA.
432 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443

stratospheric AAM signal showed in this gure is the sum of all contributions from 70 hPa up to
10 hPa [see Eq. (1)], we also investigate the correlation of the QBO in solar activity with the QBO
in each layer over the stratosphere. In an analysis for the period between 1949 and 2000, AAM at
the10 hPa level leads SA by 12 months, and such lags increase up to 76 months at the lower 50
and 70 hPa levels. Similar analysis with a monthly QBO index of three location in the equatorial
zonal winds during 19561999 (Naujokat, 1986) reveals a downward signal propagation, with
phases, at the 20, 30, 40, 50 and 70 hPa layers, increasingly lagging the solar signal by 64, 68, 71,
73, and 76 months, respectively. Such correlations approach 0.5 over this 44-year period, but
increases if we only take in account the last 30 years. Longer periods would be necessary to
reconrm the relationship with the quasi biennial oscillation characteristics (Hamilton, 1998),
though historical series are likely unobtainable.

3.2. Decadal time scales

Five decades of AAM are available from the NCEP-NCAR reanalyzes; however to study the
atmosphere during the last century, we rely on atmospheric models run with boundary sea sur-
face temperature forcing. For even lengthier analysis, we also extend our study into the historic
past by the use of a proxy climate indices of two equatorial sea surface temperature (SST) time
series (Kaplan20 and Mann5) over 17302000.

3.2.1. Decadal times scales in AAM data

To isolate decadal and longer signals in AAM, we have chosen three independent analysis
methods. Spectral analysis on unltered monthly AAM series, 19492000, reveals a powerful
decadal peak near 1011 years as well as an energetic interdecadal broad band, with a peak near
22 years. This spectrum and those of the rst and second order autoregressive (AR) models of the
detrended AAM series are represented in Fig. 4a. Both AR models highlight the statistical sig-
nicance of the decadal peak, and the AR-1 model acknowledges the signicance of the bidecadal
peak. A further wavelet analysis (Fig. 4b; Torrence and Compo, 1998) conrms the presence of
both scales throughout most of the last half century. The interdecadal oscillation passes above the
10% signicance level, according to a red noise background spectrum. Additionally, multitaper
methods (Vautard et al., 1992), shows that the decadal peak in AAM is statistically signicant
above the 95% condence level by a Fisher test, and the bidecadal with a lesser signicance at the
80% condence level. We discuss further analysis of the statistically-signicant decadal peak, and
leave the more weakly signicant bidecadal one for later study.
Using Eq. (1), we computed AAM up to dierent pressure heights: 850 hPa (1.5 km), 500 hPa
(5.5 km), 200 hPa (12 km), 100 hPa (16 km) over the troposphere, and up to 50 hPa (20 km) and
10 hPa (33 km) in the stratosphere. From the spectrum performed on each of the time series, we
derive spectral characteristics (power and phase) of the decadal peak. The decadal cycle in AAM
results from a major contribution of the troposphere (up to 100 hPa) and a signicant participa-
tion (up to 20% of the variance) of the stratosphere (10010 hPa) (Table 1). The calculations
within the lower tropospheric heights (850, 500 hPa) lead those computed over upper heights,
while the troposphere (100 hPa) leads the stratospheric layers by roughly 3 months (above 100
hPa). Although phase shifts between AAM series are small compared to the decadal period, the
mode propagates slightly upwards. Our alternative multichannel singular spectrum analysis on
 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443 433

zonal wind in vertical layers (see Abarca del Rio et al., 2000) also conrms such upward propa-
gation from the surface throughout the troposphere and into the stratosphere.
To quantify the common signal between solar activity and atmospheric angular momentum, we
applied a co-herency applied coherency spectrum (Fig. 4c; Hinich and Clay, 1968) to the AAM
and solar activity series. Results indicate signicant relationship at decadal time scales with sta-
tistical signicance (Julian, 1974) for the 95% limit (P=0.63). The phase spectrum shows no
signicant phase lead or lag between the series. The coherency spectrum suggests as well that the
series are unrelated at other low frequencies, such as the bidecadal peak in AAM. A wavelet
coherency approach (Torrence and Webster, 1999) also conrmed these results.
When the series is subjected to low-pass ltering with a cut-o at 8 years (Fig. 5a), a rela-
tionship between solar activity and AAM at decadal time scales emerges. The maximum corre-
lation between the two series is 0.74, solar activity leading LOD by 2-months.With these series

Fig. 4. Decadal variability in AAM. (a) Power spectra of AAM (solid), AR1 (dashed) and AR2 (solid grey) models. (b)
Wavelet power spectrum. Contour levels are chosen so that 75%, 50%, 25% and 5% of the wavelet power is above each level,
respectively. Black contour is the 10% signicance level, using a red noise (autoregressive lag=0.98) background spectrum. (c)
Square Coherency (solid line) and phase shift (dashed grey line) of the AAM and SA for periods from 2 to 30 years.
434 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443

Fig. 4. (continued)

Table 1
Characteristics of the decadal cycle in the AAM series integrated up to dierent pressures heights (hPa) in the
atmosphere

Height Amp (%) Lag (m)

10 100 0
50 93.2 2.5
100 80.6 3
200 60.4 3.5
500 31.6 4
850 6.3 4.5

The rst column presents the upper atmospheric height integrated. The second and third columns, the variance (in
percentage), and the phase (in months), with respect to the AAM series integrated up to 10 hPa.

band-pass ltered between 8 and 16 years (Fig. 5b), the relationship at decadal time scales is
further claried. The AAM signal shows decadal uctuations of about 0.07 ms, with peaks close
to the years 1960, 1970, 1980 and 1990, almost coincident with those of the solar cycles. The
correlation between decadal cycles increases to 0.91 with a 2-month lag (solar input leading).
When the AAM and SA series of Figs. 5a and b are directly subtracted (5a5b), separating lower
time scales, the AAM series displays interdecadal uctuations of about 0.02 ms, with peaks near
1960 and 1980, revealing a periodicity of about 20 years (in grey in Fig. 5b), agreeing therefore
with the above spectral analysis (Fig. 4a and b).
 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443 435

Fig. 5. Comparison of SA (solid line) and AAM (dashed). (a) Low pass ltered (> 8 years). (b) Band pass ltered
(816 years) and low pass ltered ( > 16 years, in grey).

Fig. 6. Top: Comparison of ensemble averages of simulations of AAM by NCAR (dashed) and UKMO (solid) and the
SA (bold solid), band pass ltered between 8 and 13 years since 1870. Bottom: Similar but for amplitude of the annual
cycle.

This gure conrms that although AAM and the solar activity series present a related varia-
bility at decadal times scales, they appear to be unrelated at lower frequencies.
We take advantage of the long series of AAM from the model runs of both NCAR (Kiehl et al.,
1998) and the Hadley Centre (Rosen and Salstein, 2000). We applied the same technique than above for
ltering the raw data, i.e. resulting in a 813 years band pass ltering. In this gure, the solar decadal
cycle and the decadal cycle in AAM agree only over the last decades (Fig. 6a). Moreover, the amplitude
of the annual cycle of AAM (Fig. 6b) appears to be well related to the variability in the solar forcing.

3.2.2. Decadal time scales in sea surface temperature (SST) elds

Historical values of SA are compared to the lengthy nearly-three century long series in sea
temperature Kaplan20 and Mann5 time series (Fig. 7a and b, respectively). We made sure to limit
436 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443

Fig. 7. Comparison of the decadal variability in SA (grey line) with the decadal variability seen in the SST Kaplan20
time series (8a), and the historical equatorial time series (Mann5) of SST elds from Mann et al. (8b).

the decadal band to the period between 8 and 13 years rather than longer as performed in White et al.
(1997) to avoid the El Nino-related band at 1415 years (Parker et al., 1994). The decadal variability
of the Kaplan20 series was compared successfully, moreover with that in the Kaplan10 and Kaplan5
series (see Section 2.3), assessing the coherence of the equatorial SST elds at these bands.
The phase variability of the two series (Kaplan20 and Mann5) agree in the period prior to 1950.
The two SST series match well, lagging or in phase with SA over much of the 20th century since
1930. However they are out of phase with SA for some periods, like the turn of the 19th to 20th
century. The analysis of SST in all ocean basin times series (not shown) conrms the disagree-
ment found at the turn of the last century. Another clear agreement with the decadal solar cycle
found is found over the middle of the 18th century (17301770) in the historical (Mann50 s)
equatorial SST time series, and conrmed by all the ocean basin (Pacic, Atlantic and Indian
oceans) times series (not shown). As well the agreement over the last 50 years can be attributed
for the most to the equatorial Indian and east Pacic ocean (not shown).
Similarly, decadal variability in AAM is linked to surface temperature (marine + land) time
series, determined from Mann et al. (2000) over 1700 to 2000 at their decadal time scales with the
Schwabe cycle (Fig. 8). This series conrms results found with SST and AAM, i.e., the decadal
 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443 437

Fig. 8. Comparison of the decadal variability in SA (grey line) and that in the global (marine+land) surface tem-
perature (GT) index by Mann et al. (2000).

variability in those climate indices agrees only coincidentally with the decadal oscillation
(Schwabe cycle) in solar activity. Temperatures in turn may be linked to AAM through the dis-
tribution of the zonal winds responding to the meridional temperature gradient.

3.2.3. Analysis of Length of day variability

Independently, Challinor (1971), Vondrak (1977) and then Currie (1980) found, over dierent
time spans a relationship between the Schwabe cycle in SA and the 1011 year in LOD. However,
the dierences in the phase shift found by the authors prevent any denitive conclusion. Currie
found a three years lag, whereas Vondrak found a relationship much closer to the one found in
the present study between AAM and SA. In Fig. 9 are displayed the SA and LOD as obtained by
a singular spectrum analysis with an embedding dimension of 13 years. The LOD and SA signal
appears to be correlated over much of the last 40 years of the 20th century, in phase opposition

Fig. 9. Comparison of Schwabe cycle in SA and decadal cycle in LOD over 17002000, as obtained by SSA with an
embedding dimension of 13 years.
438 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443

about the turn of 19th to the 20th centuries and out of phase before. Determining such relation-
ships with LOD data before 1950, particularly before 1920 is dicult due to the lower precision
then (Jordi et al., 1994). In particular, the variable evolution of phase shift between the Schwabe
cycle and the climatic decadal cycle, seen in number of surface parameters presented here is
maintained.

4. Discussion and conclusion

Atmospheric angular momentum (AAM) as a useful indicator of the whole atmospheric

dynamic state, allows an unprecedented opportunity to test and understand solar-atmosphere
connections. The results obtained in this research can be subdivided in two sections: interannual
and decadal times scales.

4.1. Interannual times scales

At interannual time scales, we extended to AAM during 19492000, the relationship between
the decadal cycle in solar activity (the so called Schwabe cycle) found by other authors and
other parameters, and the period of the quasi biennial oscillation (QBO) over the stratosphere,
here in the stratospheric part of atmospheric angular momentum (the s-QBO). A deeper analy-
sis shows that the lag between those decadal oscillations is not constant and varies slightly from
about 3 years in 1965 to 5 years in 1995. It is also possible that the internal tropospheric climate
variability plays a role too in the quasi-biennial period variability, though less likely, according
to our results. Additionally, these phase lags could be contaminated by other sources of internal
variability, such as volcanism (Robock, 1978) which may aect stratospheric variability, or the
proximity of a decadal oscillation in ENSO (Hanson et al., 1989; Dovgalyuk and Klimenko,
1996). Nevertheless, the amplitude variability of the stratospheric QBO in AAM is closer too to
the decadal scales seen in ENSO (1314 years) (Hanson et al., 1986; Dovgalyuk and Klimenko,
1996).
The quasi biennial cycle in solar activity is also interesting. QBO in the stratospheric compo-
nent of AAM as well as that of the index of the stratospheric winds (Naujokat, 1986) presents a
phase variability lagged by 6 years (73 months) with regards to the QB cycle present in solar
activity, explaining a result found by Djurovic and Paquet (1993) regarding a correlation between
LOD and solar activity at this band. However the weak amplitude of this signal in solar activity
(when compared to the more powerful Schwabe cycle) as well as the time dependence of the
association found (the signals over the stratosphere are in phase only when the amplitude of the
solar QB cycle is strong) prevents any denitive conclusion. It is dicult to believe that a pow-
erful decadal cycle may not imprint the decadal variability in climate, and it is the case for a
weaker QBO. An interesting possibility is to consider the Schwabe cycle to control the phase of
the quasi biennial cycle in solar activity and in the stratosphere. Therefore, both relationships
could support each other independently. The 6-year lag between both quasi biennial oscillations
could be interpreted as an indication that the association between solar variability and climate is
lagged. Consequently, we might consider that the atmosphere is imprinted by solar activity at this
scale. This possible association needs further investigations.
 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443 439

Conversely, the quasibiennial oscillation and the tri-quadrennial oscillation over the tropo-
sphere do not appear to be related with solar activity, either directly or modulated by decadal
variability and comparable phase variability. In the case of the quasi-biennial cycle, it appears
that it is somewhat correlated with the quasi biennial oscillation in solar activity (QBO-SA) over
the last 50 years, though leading SA through part of the last century. It is fortunate here that an
monthly index of the southern oscillation went back as far as 1866, which allows us to test this lag
through times. Unfortunately, determination of stratospheric variability typically requires in situ
data, and thus is not extant prior to the rocket age as well as not much information is available
prior to the International Geophysical Year, 1958.

4.2. Decadal times scales

Relative AAM, based on zonal wind reanalysis of NCEP/NCAR from 1949 to 2000, integrated
up to 10 hPa (99% of the total mass of the atmosphere), exhibits a signicant decadal period of
1012 years and an interdecadal oscillations of 22 years, with somewhat less statistical sig-
nicance. Its variance, roughly 0.07 and 0.03 ms, respectively, are important for the global
angular momentum of the Earth and should be further investigated.
The decadal cycle in AAM is shown here to be due to the whole atmosphere in both magnitude
and phase, with the troposphere contributing up to 80% of the total power. We also showed that
this decadal cycle can originate in surface processes, as the troposphere leads the stratosphere.
The decadal cycle in the global AAM presents no signicant phase shift (2 months in average)
with the solar decadal cycle.
At the turn of the 20th century, decadal variability in AAM as given by the NCAR and Hadley
Centre atmospheric model runs, as well as in LOD, appears to be out of phase with the Schwabe
cycle, and thus in phase with each other. It is interesting to note here, that although the quality of
LOD data is considerably lower at the early part of the 20th century, it could kept well the phase
of the atmospheric signature.
On the other hand, Lau and Weng (1995) as White et al. (1997), nd signicant solar-related signals in
equatorial SST over the entire century. The discrepancy seen between their SST signals and our AAM
series, given particularly that the UKMO and NCAR model runs from which is issued the AAM
investigated here were forced by SST uxes, lead us to investigate decadal variability in SST. The two
SST series, the Kaplan20 (18562000) and Mann5 (17301980) indicate a lag with solar activity since
1930 but is out of phase prior to about 1900; however, partial agreement was present during the mid 18th
century. We extended the comparison to decadal variability in the global surface (land + marine) tem-
perature (GT) reconstruction by Mann et al. (2000) back to 1700 and found agreement with the
above; series of GT are similar to those of SST regarding agreement with the solar signal. It
appears therefore that independent measurements (SST, GT, AAM, LOD) conrms the out of
phase lag of the climatic decadal variability at the turn of the 19th to the 20th century, as well as
the partial agreement over the last 50 years. Based on this historical analysis of decadal variability
in climate and geodetic variables, it is possible to say that the particular relationship found since
at least 50 years between SA and equatorial SST is most probably coincidental.
Nonetheless, note that a partial relationship over the last few decades and phase opposition at the
turn of the century, which may agree well with a relationship found at decadal times scales
between solar activity, surface air temperature changes (Nitta and Yoshimura, 1993; Parker et al.,
440 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443

1994), decadal variability in SST, AAM and LOD, can be also very well explained by the presence of
very close independent decadal periods in solar activity and climate (for example, 11 years and 10.2
years).
We may note here that the failure of decadal atmospheric variations to remain coherent with
the solar cycle prior to around 1900 does not preclude such agreement since that time. As dis-
cussed by Feynmann and Crooker (1978), the yearly means of the geomagnetic index Aa, which is
representative of solar wind velocity and magnetic eld disturbances were between a half and a
third of its strength during the last three decades when compared to that at the beginning of the
century. With weaker forcing by the solar wind the phase lock to the solar cycle of internal decadal
variability could have been lost. Also, reconstruction of past solar irradiance (Hoyt and Schatten,
1997) shows an increase since at least 1940, with the largest irradiance ever measured, a likely
explanation for the current association. Moreover, since decadal atmospheric variability may not
uniquely identied with the 11-year solar cycle, a lack of full agreement with the solar forcing even
more recently does not exclude its impact on the climate and momentum-related indices.
However, whatever the source of the decadal cycle in AAM may be, internal or external, this
oscillation in the global atmosphere (troposphere and stratosphere) is suggested here to be gen-
erated by upward propagation of surface atmospheric modes, from the surface throughout the
troposphere through the stratosphere. Because AAM represents the dynamic state of the global
atmosphere, decadal vacillations seen in dierent parameters over the stratosphere (geopotential
height, temperature, zonal winds, clouds, see Section 1) are most probably associated with these
dynamics, and show the inuence of surface processes within the troposphere rather than an
external forcing directly into the middle atmosphere. Also, because the long-term analysis per-
formed here in either proxy, reconstructed or measurements associated with tropical SST or global
temperature do not show a full association with the Schwabe cycle, such decadal stratospheric
vacillations associated with the solar cycle may not have been evident in past times. However, the
strength of the stratospheric QBO indicates an atmospheric role, through historical tropical surface
temperature and other measurements. The present results indicate the need for better understanding
of atmospheric dynamics at decadal time scales. It seems that the coming years will be fruitful in this
regard, given the advent of extended and improved atmospheric and solar data.

Acknowledgements

The authors are grateful to two referees whose useful comments helped improve the original
version of the paper. The National Center for Atmospheric Research is sponsored by the US
National Science Foundation. The work is partially supported by the US NASA Solid Earth and
Natural Hazards Program.

References

Abarca del Rio, R., Gambis, D., Salstein, D.A., 2000. Interannual signals in length of day and atmospheric angular
momentum. Ann. Geophys. 18, 347364.
Abarca del Rio, R., 1997. La rotation de la Terre: Etude du cycle annuel et de la variabilite basse frequence. Etude
climatique de la rotation Terrestre. PhD thesis, Universite Paul Sabatier.
 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443 441

Akioka, M., et al., 1987. The 17-month periodicity of sunspot activity. Solar. Phys. 112, 313316.
Arnold, N.F., Robinson, T.R., 1998. Solar cycle changes to planetary wave propagation and their inuence on the
middle atmosphere circulation. Ann. Geophys. 16, 6976.
Baldwin, M.P., Gray, L.J., Dunkerton, T.J., Hamilton, K., Haynes, P.H., Randel, W.J., Holton, J.R., Alexander, M.J.,
Hirota, I., Horinouchi, T., Jones, D.B.A., Kinnersley, J.S., Marquardt, C., Sato, K., Takahashi, M., 2001. The
Quasi-biennial oscillation. Reviews of Geophysics 39 (2), 179229.
Bao, S., Zhang, H., 1998. Patterns of current helicity for the solar cycle 22. Astrophys. J. 496, 4346.
Barnes, R.T.H., Hide, R., White, A.A., Wilson, C.A., 1983. Atmospheric angular momentum uctuations, length of
day changes and polar motion. Proc. R. Soc. Lon. 387, 3173.
Benevolenskaya, E.E., 1998. A model of the double magnetic cycle of the sun. Astrophys. J. 59, 4952.
Berson, F.A., Kulkarni, R.N., 1968. Sunspot cycle and the quasibiennial oscillation. Nature 217, 11331134.
Bertrand, C., van Ypersele, J.-P., Berger, A., 1999. Volcanic and solar impacts on climate since 1700. Clim. Dyn. 15,
355367.
Black, R.X., Salstein, D.A., Rosen, R.D., 1996. Interannual modes of variability in atmospheric angular momentum. J.
Clim. 9, 28342849.
Challinor, R.A., 1971. Variations in the rate of rotation of the Earth. Science 172, 10221024.
Chen, T.-C., Tribbia, J.J., Yen, M.C., 1996. Interannual variation of global atmospheric angular momentum. J. Atm.
Sci. 53, 28522857.
Crowley, T.J., Kim, K.-Y., 1996. Comparison of proxy records of climate change and solar forcing. Geophys. Res.
Lett. 23, 359362.
Crowley, T.J., 2001. Causes of climate change over the past 1000 years. Science 289, 270277.
Cubasch, U., Voss, R., Hegerl, G.C., Waszkewitz, J., Crowley, T.J., 1997. Simulation of the inuence of solar radiation
variations on the global climate with an ocean-atmosphere general circulation model. Clim. Dyn. 13, 757767.
Currie, R.G., 1980. Detection of the 11-yr sunspot cycle signal in Earth rotation. Geophys. J. R. Astr. Soc. 61, 131140.
Diamantides, N.D., 1998. Long-term solar activity and terrestrial connections. Part II: at the beckon of the sun? Ann.
Geophys. 16, 492509.
Dickey, J.O., Marcus, S.L., Hide, R., 1992. Global propagation of interannual uctuations in atmospheric angular
momentum. Nature 357, 484488.
Dickey, J.O., Gegout, P., Marcus, S.L., 1999. Earth-atmosphere angular momentum exchange and ENSO: the rota-
tional signature of the 199798 event. Geophys. Res. Lett. 26, 24772488.
Djurovic, D., Paquet, P., 1993. Quasi-biennial oscillation in green corona activity and Earth rotation. Astron. Astro-
phys. 277, 669676.
Djurovic, D., Paquet, P., 1996. The common oscillations of solar activity, the geomagnetic eld, and the Earth rota-
tion. Solar. Phys. 167, 427439.
Dovgalyuk, V.V., Klimenko, V.V., 1996. On long term variations in the intensity of El Nino occurrences. Geophys.
Res. Lett. 23, 37713774.
Eddy, J.A., 1976. The Maunder Minimum. Science 192, 11891202.
Eubanks, T.M., 1993. Variations in the orientation of the Earth. In Contributions of Space Geodesy to geodynamics,
edited by S.D.E., and D.L. Turcotte, 24, pp. 154, American Geophysical Union, Washington, DC.
Feynman, J., Crooker, N.U., 1978. The solar cycle at the turn of the century. Nature 275, 626.
Friis-Christensen, E., Lassen, K., 1991. Length of the solar cycle: an indicator of solar activity closely associated with
climate. Science 254, 698700.
IERS Annual Report for 1999, Observatoire de Paris, Gambis (Ed.).
Gierens, K., Ponater, M., 1997. Comment on Variation of cosmic ray ux and global cloud coveragea missing link
in solar-climate relationships by H. Svensmark and E. Friis-Christensen. J. Atmos. Solar-Terr. Phys. 61, 795797.
Gorney, D.J., 1990. Solar cycle eects on the near-Earth space environment. Rev. Geophys. 28, 315336.
Gross, R.S., 2001. A combined length-of-day series spanning 18321997: LUNAR97. Phys. Earth Planet. Inter. 123,
6576.
Hamilton, K., 1990. A look at the recently proposed solar-QBO-weather relationship. J. Clim. 3, 497503.
Hamilton, K., 1998. Observations of tropical stratospheric winds before world war II. Bull. Amer. Meteor. Soc. 79,
13671371.
442 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443

Hanson, K., Brier, G.W., Maul, G.A., 1989. Evidence of signicant non random behavior in the recurrence of strong
El Nino between 1525 and 1988. Geophys. Res. Lett. 16, 11811184.
Hinich, M.J., Clay, C.S., 1968. The application of the discrete Fourier transform in the estimation of power spectra,
coherence, and bispectra of geophysical data. Rev. Geophys. 6, 347363.
Hopfner, J., 1998. Seasonal variations in length of day and atmospheric angular momentum. Geophys. J. Int. 135, 407437.
Hopfner, J., 2001. Atmospheric, oceanic and hydrological contributions to seasonal variations in length of day. J.
Geod. 75, 137150.
Hoyt, D.V., Schatten, K.H., 1997. The Role of the Sun in Climate Change. Oxford University Press, New York.
IPCC (Intergovernmental Panel on Climate Change), 2001. Climate change. In: Houghton, G.J., Jenkins, G.J.,
Ephraums, J.J. (Eds.), The IPCC scientic assessment. Cambridge University Press, Cambridge, p. 364.
Jordi, C., Morrison, L.V., Rosen, R.D., Salstein, D.A., Rosello, G., 1994. Fluctuations in the earths rotation since
1830 from high resolution astronomical data. Geophys. J. Int. 117, 811818.
Julian, P.R., 1974. Comments on the determination of signicance of the coherence statistic. J. Atm. Res. 32, 836837.
Kalnay, E., Kanahitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, H., Saha, S., White, S., Woollen,
J., Zhu, Y., Chelliah, M., Ebisuzaki, W., Higgins, W., Jandwiak, J., Mo, K.C., Ropelewski, C., Wang, J., Leetma,
A., Reynolds, R., Jenne, R., Joseph, D., 1996. The NCEP/NCAR 40-year reanalysis project. Bull. Amer. Meteor.
Soc. 77, 437471.
Kaplan, A., Cane, M., Kushnir, Y., Clement, A., Blumenthal, B., Rajagopalan, B., 1998. Analyses of global sea surface
temperature 18561991. J. Geophys. Res. 103, 1856718589.
Kiehl, J.K., Hack, J.J., Bonan, G.B., Boville, B.A., Williamson, D.L., Rasch, P.J., 1998. The National Center for
Atmospheric Research Community Climate Model: CCM3. J. Clim 11, 11511178.
Kishcha, P.V., Dmitrieva, I.V., Obridko, V.N., 1999. Long-term variations of the solar-geomagnetic correlation, total
solar irradiance, and the northern hemispheric temperature (18681997). J. Atm. Terr. Phys. 61, 799808.
Kodera, K., 1991. The solar and equatorial QBO inuences on the stratospheric circulation during early northern
winter. Geophys. Res. Lett. 18, 10231026.
Labitzke, K., 1987. Sunspots, the QBO, and the stratospheric temperature in the north polar region. Geophys. Res.
Lett. 14, 535537.
Labitzke, K., van Loon, H., 1990. Association between the 11-year solar cycle, the quasi-biennial oscillation and the
atmosphere: a summary of recent work. Phil. Trans. Roy. Soc. London., A 330, 577589.
Labitzke, K., van Loon, H., 1997. The signal of the 11-year sunspot cycle in the upper troposphere-lower stratosphere.
Space Science Reviews 80, 393410.
Lau, K.-M., Weng, H., 1995. Climate signal detection using wavelet transform: how to make a time series sing. Bull.
Am. Meteor. Soc. 76, 23912402.
Lean, J., Rind, D., 1999. Evaluating sun-climate relationships since the little ice age. J. Atm. Terr. Phys. 61, 2535.
Liao, D.C., Greiner-Mai, H., 1999. A new DELTA LOD series in monthly intervals (1892.0-1997.0) and its compar-
ison with other geophysical results. J. Geodesy 73, 466477.
Mann, M.E., Gille, E., Bradley, R.S., Hughes, M.K., Overpeck, J., Keimig, F.T., Gross, W., 2000. Global temperature
patterns in past centuries : an interactive presentation. Earth Interactions 4, 4, 128.
Marcus, S.L., Chao, Y., Dickey, J.O., Gegout, P., 1998. Detection and modeling of non tidal oceanic eects on earths
rotation rate. Science 281, 16561659.
Meehl, G.A., Washington, W.M., Wigley, T.M.L, Arblaster, J.M., Dai, A., 2003. Solar and greenhouse gas forcing and
climate response in the twentieth century. J. Climate 16, 426444.
Michaelsen, J., 1986. Long period uctuations in El Nino amplitude and frequency reconstructed from tree-rings. In:
Peterson, D.H. (Ed.), Aspects of Climate Variability in the Pacic and Western Americas, Geophysical Monograph
55. American Geophysical Union, Washington DC, pp. 6974.
Naito, Y., Hirota, I., 1997. Interannual variability of the northern winter stratospheric circulation related to the QBO
and the solar cycle. J. Meteor. Soc. Japan 75, 925937.
Naujokat, B., 1986. An update of the observed quasi-biennial oscillation of the stratospheric winds over the tropics. J.
Atmos. Sci. 43, 18731877.
Nesme-Ribes, E., Thuillier, G., 2000. Histoire solaire et climatique, Ed. Belin, Paris, Coll Pour la Science (Scientic
American).
 R. Abarca del Rio et al. / Journal of Geodynamics 36 (2003) 423443 443

Nitta, T., Yoshimura, J., 1993. Trends and interannual and interdecadal variations of global land surface air tem-
perature. J. Meteor. Soc. Japan 67, 375383.
Pap, J.P., Frohlich, C., 1999. Total solar irradiance variations. J. Atm. Terr. Phys. 61, 1524.
Parker, D.E., Jones, P.D., Folland, C.K., Bevan, A., 1994. Interdecadal changes of surface temperature since the late
nineteenth century. J. Geophys. Res. 99, 1437314399.
Parker, E.N., 2000. The physics of the sun and the gateway to the stars. Physics Today 53, 2631.
Peixoto, J.P., Oort, A.H., 1992. Physics of Climate. American Institute of Physics, New York.
Pittock, A.B., 1978. A critical look at long-term sun-weather relationships. Rev. Geophys. Space. Phys. 16, 400420.
Pittock, A.B., 1983. Solar variability, weather and climate: an update. Quart. J. Roy. Met. Soc. 109, 2357.
Quiroz, R.S., 1981. Period modulation of the stratospheric quasi-biennial oscillation. Mon. Wea. Rev. 109, 665674.
Reid, G.C., 1995. The sun-climate question: Is there a real connection ? Rev. Geophys. 33, 535538.
Riven, Y.R., Obidko, V.N., 1992. Frequency composition of multiyear variations of the magnetic eld of the sun as a
whole. Sov. Astron. (American Institute of Physics translation) 36, 557560.
Robock, A., 1978. Internally and externally caused climate changes. Amer. Meteor. Soc. 35, 11111122.
Rosen, R.D., Salstein, D.A., 1983. Variations in atmospheric angular momentum on global and regional scales and
length of day. J. Geophys. Res. 88, 54515470.
Rosen, R.D., Salstein, D.A., Wood, D.A., 1991. Zonal contributions to global momentum variations on intrasea-
sonnal through interannual time scales. J. Geophys. Res. 96, 51455151.
Rosen, R.D., Salstein, D.A., 2000. Multidecadal signals in the interannual variability of atmospheric angular momen-
tum. Clim. Dyn. 16, 693700.
Sadourny, R., 1994. Linuence du Soleil sur le climat. C.R. Acad. Sci. Paris 319, 13251342.
Salby, M.L., Shea, D.J., 1991. Correlations between solar activity and the atmosphere: an unphysical explanation. J.
Geophys. Res. 96, 2257922595.
Sakurai, K., 1979. Quasi-biennial variations of the solar neutrino ux and solar activity. Nature 278, 146148.
Sakurai, K., 1981. Quasi-biennial periodicity in the solar neutrino ux and its relation to the solar structure. Solar
Phys. 74, 3541.
Schuh, H., Schmitz-Hubsch, H., 2000. Short period variations in Earth rotation as seen by VLBI. Surveys in Geo-
physics 21, 499520.
Schwabe, H., 1843. Solar observations during 1843. Astron. Nachr. 21, 233237.
Shapiro, R., Ward, F., 1962. A neglected cycle in sunspot numbers. J. Atmos. Sci. 19, 506508.
Shindell, D.T., Rind, D., Balachandran, N., Lean, J., Lonergan, P., 1999. Solar cycle variability, ozone, and climate.
Science 284, 305308.
Sun-Climate Connections, 2000. In: Sprigg, William A. Nasa Workshop in Sun-Climate Connections: Summary
report, 31 pp., University of Arizona, Arizona.
Svensmark, H., Friis-Christensen, E., 1997. Variation of cosmic ray ux and global cloud coveragea missing link in
solar-climate relationships. J. Atmos. Solar-Terr. Phys. 59, 12251232.
Tett, S.F.B., Stott, P.A., Allen, M.A., Ingram, W.J., Mitchell, J.F.B., 1999. Causes of twentieth-century temperature
change near the Earths surface. Nature 399, 569572.
Torrence, C., Compo, G.P., 1998. A practical guide to wavelet analysis. Bull. Amer. Meteor. Soc. 123, 309336.
Torrence, C., Webster, P.J., 1999. Interdecadal changes in the ENSO-Monsoon system. J. Clim. 12, 26792690.
Vautard, R., Yiou, P., Ghil, M., 1992. Singular spectrum analysis: a toolkit for chaotic time series. Physica D 38, 74
95.
Vondrak, J., 1977. The rotation of the Earth between 1955.5 and 1976.5. Studia geoph. et geod 21, 107117.
Wang, B., Wang, Y., 1996. Temporal structure of the southern oscillation as revealed by waveform and wavelet
analysis. J. Clim. 9, 15861598.
White, A.A., 1989. A relationship between energy and angular momentum conservation on dynamical models. J. Atm.
Sci. 46, 18551860.
White, W.B., Lean, J., Cayan, D.R., Dettinger, M.D., 1997. Response of global upper ocean temperature to changing
solar irradiance. J. Geophys. Res. 102, 32553266.